To tune query performance you can index JSON fields in several ways, store their values in the In-Memory Column Store (IM column store), or expose them as non-JSON data using materialized views.

Chapters:

• Overview of Performance Tuning for JSON
  Which performance-tuning approaches you take depend on the needs of your application. Some use cases and recommended solutions are outlined here.
• Indexes for JSON Data
  You can index JSON data as you would any data of the type you use to store it. In addition, you can define a JSON search index, which is useful for both ad hoc structural queries and full-text queries.
• In-Memory JSON Data
  A column of JSON data can be stored in the In-Memory Column Store (IM column store) to improve query performance.